SEO v. SEA  |  SMART Campaigns  |  Search Analytics
	Most companies test combinations of various attributes of their advertising (copy, design, price, etc.) before launching a campaign. This traditional method of testing all possibilities is not feasible for search engine marketing because of the large number of advertisements typically used in such campaigns. However, thanks to “experimental design,” a powerful statistical technique, you can test just a few combinations and interpolate the remainder to find the best combination. Traditional Testing v. Experimental Design Suppose you want to test two different search engine ads, two different landing pages, and two different prices. With traditional testing, you would run eight tests—one for each combination of search engine ads, landing pages, and prices. Most companies can afford to run eight different tests, but this traditional method of testing all possibilities can quickly become too expensive and time-consuming for even the world’s largest companies. Regarding the above example, if you instead want to test a more realistic eight search engine ads, two landing pages, and four prices, you would need to run 64 tests using traditional methods. By the time you finished testing, the product you’re trying to sell might reach the end of its life cycle! Additionally, traditional testing may tell you that the best combination is search engine ad 1, landing page 2, and price 2, but it doesn’t say anything about these attributes individually. Furthermore, items like search engine ads and landing pages have multiple attributes of their own (title, call to action, layout, etc.). In the above example, breaking down the eight search engine ads into two components and the two landing pages into three components would require 384 tests. Experimental design makes the impossible possible. Originally developed for hard sciences like chemistry and physics, experimental design takes much of the guesswork out of search engine marketing. Using “logistic regression analysis,” experimental design enables you to run the equivalent of 64 or even 384 tests. Before you can plug in the numbers, however, you need to design the experiment—a task best placed in the hands of someone with expertise in statistics and search engine advertising. Only a properly designed experiment can produce reliable results. You're Busy. We Can Help. Experimental design makes SMART search engine campaigns SMARTER by adding “Efficacy” and “Response” to the mix. We pioneered the use of this statistical technique for optimizing search engine advertising campaigns. By putting our expertise to work for your company, you can gain a significant advantage over your competitors, even those that also use search engine advertising. Contact us to discuss your next search engine marketing campaign